Llama 3.1

Old

Meta

•

Open Source

•

70B

# 50

Released

Jul 23, 2024

# 9

Knowledge Cutoff

Dec 23

# 7

Context Length

128K

Benchmarks

# 53

Code LiveBench

32.7%

# 41

Code LMArena

1209

# 50

GPQA Diamond

43.0%

# 39

Reason LiveBench

40.7%

# 42

ELO LMArena

1242

# 42

Math LiveBench

34.4%

# 36

MATH

60.0%

# 29

Human Eval

80.5%

# 20

Aider Old

58.6%

# 34

MMLU Pro

66.4%

# 14

MMLU

86.0%

# 24

Halluc. Hughes

5.0%

# 33

IF LiveBench

69.0%

# 35

Data LiveBench

49.1%

# 40

Lang LiveBench

34.6%

# 38

Avg LiveBench

50.9%

# 6

IF Evaluation

87.5%

Pricing

# 22

Input Cost /M

$1

# 21

Output Cost /M

$3